Inappropriate medication prescribing is an important problem in the elderly, but is difficult to measure. As part of a randomized controlled trial to evaluate the effectiveness of a pharmacist intervention among elderly veterans using many medications, we developed the Medication Appropriateness Index (MAI), which uses implicit criteria to measure elements of appropriate prescribing. This paper describes the development and validation of a weighting scheme used to produce a single summated MAI score per medication. Using this weighting scheme, two clinical pharmacists rated 105 medications prescribed to 10 elderly veterans from a general medicine clinic. The summated score demonstrated acceptable reliability (intraclass correlation co-efficient = 0.74). In addition, the summated MAI adequately reflected the putative heterogeneity in prescribing appropriateness among 1644 medications prescribed to 208 elderly veterans in the same general medicine clinic. These data support the content validity of the summated MAI. The MAI appears to be a relatively reliable, valid measure of prescribing appropriateness and may be useful for research studies, quality improvement programs, and patient care.